Brain arteriovenous malformations (AVMs) are an important cause of intracranial hemorrhage (ICH) in young adults. The etiology and pathogenesis are unknown and better understanding of the signaling pathways influencing disease susceptibility and clinical progression is needed to optimize patient care. In the current funding period, we have made several novel and important discoveries, reporting the first-ever genetic associations with AVM susceptibility and intracranial hemorrhage (ICH). These findings set the stage for work during the next period, during which we will study the influence of genetic variation on disease susceptibility in a focused and comprehensive fashion using a large existing DMA collection from AVM cases (n=647) and expected new cases to be enrolled (n=370) over the course of the project, yielding a total sample size of 1017 cases. Aim 1 is a Genome-wide association (GWAS) study to identify novel candidate variants associated with AVM, using the Affymetrix SNP Array 6.0, containing -1 million SNPs and ~1 million probes for copy number variation (CNV). The primary analysis will compare AVM cases to healthy controls in our largest sub-group, Caucasians (520 cases and 643 controls), using novel genomic methods to genetically match cases and controls and adjust for population stratification using principal components analysis. Secondary analysis will evaluate whether CNV is associated with AVM susceptibility. Aim 2 will follow-up candidate variants associated with susceptibility through fine-mapping of approximately 20 target regions containing the most highly significant SNPs and/or CNVs from GWAS analysis through a variety of standard methods. Our top-associated candidate SNPs will be validated in AVM cases and controls from the University of Bonn (Germany) and University of Utrecht (Netherlands). Additional analyses will test other race-ethnic groups and time-to-subsequent hemorrhage in AVM cases. Further, biological function of AVM-associated variants will be tested using standard molecular techniques. Aim 3 will be exploration of somatic variation as a genetic mechanism in AVM using CNV analysis to compare genomic DNA to lesional DNA isolated from surgical specimens by examining loss of heterozygosity or allelic imbalance. In 100 patients, we will genotype DNA from vascular tissue that is dissected from the AVM nidus removed at surgery with Affymetrix 6.0 arrays, and compare CNVs genome-wide between germline DNA and lesion DNA, and follow-up will include use of subcloning and deep re-sequencing.